The startup says revenue grew 50-fold in two years and bookings passed $150 million. It declined to give an annual run rate, and its bookings figure includes multiyear contracts.
Ema, a startup that uses teams of AI agents to automate corporate work in human resources, IT and finance, has raised $77 million in a Series B round and more than quadrupled its valuation since its last raise in 2024.
Bengaluru-based venture firm Creaegis led the round, with existing investors Accel, Section 32 and Prosus increasing their stakes. The round was entirely primary equity, with no debt or secondary share sales, bringing Ema's total funding to $140 million. The company declined to disclose its new valuation.
Founded in 2023 by former Google and Coinbase executive Surojit Chatterjee and former Okta executive Souvik Sen, Ema sells what it calls "AI employees," systems that coordinate multiple agents to carry out multistep business processes across a company's existing applications.
Chatterjee's pitch goes further than automation. He says Ema first wraps around a customer's existing software and then lets it reduce dependence on those products over time. "Many of our customers are already on the way to replace [large SaaS applications] completely, removing dependency on them, because they are mostly becoming like a database," he said.
The company's reported metrics are strong. Ema says it has more than 50 active enterprise deals, over 1 million active enterprise users and more than 5 million actions and queries handled. Customers include NTT DATA, Hitachi, ADP, PwC, Google, KPMG, Wipro and Microsoft. It says revenue has grown 50-fold over two years, more than 90% of customers have expanded beyond their first use case, net dollar retention is about 180% and gross margins are close to 80%.
The bookings figure needs careful reading. Ema says revenue bookings have passed $150 million, but Chatterjee said that total includes the full value of multiyear contracts, including two- and three-year deals, rather than annual recurring revenue. He declined to give the company's current annualized revenue run rate. Because the total spreads across contracts of two and three years, it suggests annual revenue well below the headline bookings figure.
Ema's pricing model is also unusual. It does not charge per software seat or per AI token consumed. Instead, pricing is tied to completed tasks and business outcomes, a structure that links revenue directly to whether the agents do their jobs.
The competitive field is crowded. Anthropic has expanded efforts to bring Claude into companies' core financial and legal operations, and OpenAI has built teams of engineers who work alongside customers. Chatterjee does not see the frontier labs as direct rivals. Ema draws on more than 150 models and focuses on integrations and domain knowledge. "Progress in frontier models is actually very beneficial to us," he said.
Most of the new money will go into sales and marketing, and Ema plans to expand into Asia-Pacific, South America and parts of the Middle East over the next year. The company has nearly 200 employees.
The larger question for investors is the one Chatterjee raised himself. If AI agents can take over work companies now pay software vendors and IT services firms to do, spending shifts away from incumbents. Ema's growth offers one early data point. A disclosed annual revenue figure would make it a much clearer one.